Adaptive Radiotherapy for HN Cancer: Technical Aspects

Di Yan, D.Sc, FAAPM

Beaumont Health System Michigan, USA

Status of ART in the Clinic

Clinical rationales of ART for HN caner

ART technologies & implementation in imaging, feedback & planning modification

Practical issues of ART in clinical operation

Learning Objectives

Significant normal tissue toxicities have been caused by the large treatment volume, and organ over dose during the treatment delivery due to

• Patient/organ position & volume variations

• Cavity shape variation (induced hot-spot on mucosa)

• Neck and shoulder flexing in treatment setup

• Shrinkage of large tumor & edema resolving

Can online anatomical image (CBCT, CT, MRI) guided ART reduce normal tissue toxicities?

HN Cancer ART: Clinical Rationales

Patie

nt/O

rgan P

ositio

n V

aria

tion

Org

an V

olu

me/D

ose V

aria

tions

Organ Dose Variations during the Treatment

Left Poratid: Cumulative Dose (Dmean) Variation

0

2

4

6

8

10

12

14

-15% -10% -5% 0% 5% 10% 15% 20% 25%

%Dose Variation (Delived - Planned)

No

of

Pa

tien

ts

Right Poratid: Cumulative Dose (Dmean) Variation

0

2

4

6

8

10

12

14

-5% 0% 5% 10% 15% 20%

%Dose Variation (Delived - Planned)

No

of

Pa

tie

nts

Cord: Cumulative Dose (Dmax) Variation

0

2

4

6

8

10

12

-6% -4% -2% 0% 2% 4% 6% 8% 10%

%Dose Variation (Delived - Planned)

No

of

Pa

tie

nts

Systematic approach to escalate or deescalate treatment dose based on spatial tumor cell bio-activities, such as

• Biological image markers to determine the most resistant tumor cells, which include

o PET; MRI: pre-treatment image, as well as the imaging of early treatment response

o spatial bio-parametric distribution in the planning objectives for dose painting

Can biological image guided (PET, MRI) ART be used to select patients, and improve their tumor control & long term survival?

HN Cancer ART: Clinical Rationales

Imaging (CBCT/CT-in-room), Feedback & Adaptation

1. Pre-treatment Simulation & Planning

Standard CT simulation & IMRT planning

0~5mm CTV-to-PTV margins & 5~7 beams

Planning CT image w/wo pre-selected bony

structures (adjacent to the target, C2-C5)

selected as the reference for daily treatment

localization & correction

Segmentation (commercial tools for auto-

segmentation), inverse planning, evaluation &

QA: 2~4 days

HN Cancer ART: Clinical Implementation

2. Daily CBCT/CT-in-room Localization & Correction

Pre-treatment CBCT/CT imaging for patient at the treatment position (~2 mins)

Bony (C2-C5) registration to the reference image by using the pre-selected bony structure (2~5 mins)

Couch translational correction (1~2 mins)

Imaging/registration/correction (commercial tools): 5~9 mins per treatment

Post-treatment image: once a week for QA purpose

HN Cancer ART: Clinical Implementation

Daily Treatment Localization

3. Daily/Weekly Treatment Evaluation and QA

Patient/organ position/volume/dose evaluation

(2~4hrs per week per patient)

Non or few commercial tools with very limited

functions at the present time can be applied for

this task

Technologies:

o CBCT-to-CT deformable image registration

o Organ position & volume variation quantification

o Daily CBCT density mapping & dose calculation

o Daily & cumulative treatment dose construction

HN Cancer ART: Clinical Implementation

Daily/Weekly Treatment Evaluation/QA

4. New CT Simulation (after the first 10 and/or 20 treatment days)

New mask if necessary

Delineate targets and ROIs on the new CT image (auto propagation from the pre-treatment plan)

The new CT image will be used in the planning modification, and as the new reference image for the rest of daily image guidance

1~2 working days depending on the level of automation in segmentation & planning

This step could be replaced using the daily CBCT directly in future

HN Cancer ART: Clinical Implementation

5. IMRT Re-planning or Adaptive Inverse Planning

Re-planning on the new CT image (1~2 days)

o on a commercial planning system

o the initial planning objectives, constraints &

weights can be used as the guidelines

Adaptive inverse planning by including all daily

CBCT images obtained during the last week,

o organ variations in the objectives of inverse

planning optimization

o Auto-planning & evaluation (1~2 days)

HN Cancer ART: Clinical Implementation

Technical Issue: Deformable Image Registration

DVF

Mesh Structure

Technical Issue: Organ Variation Characterization

Patient 2 Patient 1

0

2

4

6

8

10

0 5 10 15 20 25 30 35Fraction #

Dis

pla

ce

me

nt (m

m)

Point 1

Point 2

Patient 3

L R

Technical Issue: Treatment Dose Construction

)ˆ),((|),ˆ),(()(|)( 11 jjKjjKKN vxvxdvDvD

Delivered dose Estimated dose

Expected Treatment Dose

Technical Issue: Adaptive Inverse Planning

kNkN

kN

k

DGDG

DFMax

ˆ||

|

1

1

1

“Expected Treatment Dose” in the objective & constraints

to determine the new or modified plan

* : Expected improvement from the previous treatment is used to

determine if “accepting the plan modification”

ART vs Conventional IMRT (5mm Target Margin)

0

10

20

30

40

50

60

70

80

CTV1

CTV2

Brain

_Stem

Cord

Man

dible

Left_

Parotid

Right

_Parotid

Org

an

Cu

mu

lati

ve

Do

se

(G

y)

PreTxPlan

FinalAdap

Improvement in Gy: 9.5+6 7.2+2 6.5+10 6.7+10 7.8+3

Dmax Dmax Dmax Dmean Dmean

Conventional IMRT

Weekly adaptation

Improvement of ART vs Clinical Efforts

0%

5%

10%

15%

20%

25%

30%

35%

40%

Brain_Stem Cord Mandible LP RP

% o

f re

du

cti

on

wrt

th

e P

re-T

x C

um

. Do

se

1 Adaptive Modification 2 Adaptive Modifications 3 Adaptive Modifications

4 Adaptive Modifications 5 Adaptive Modifications

‘Daily IGRT’ vs ‘Hybrid ART’

All treatment organ doses are normalized to the baseline IMRT plan with 0 target margin

T1: Daily IGRT with 0 target

margin

T2: Daily IGRT + two weekly replanning

T3: Daily IGRT + two adaptive planning

*Dose heterogeneity in targets could be a major concern

‘Daily IGRT’ vs ‘Hybrid ART’

Segmentation: 2~3 CTs and/or daily CBCTs

Manual: ~5 hrs per image

Auto + manual editing: 10 mins ~ 3 hrs per image

Planning: 2~3 times

Manual: 6 hrs per plan

Auto + manual modification: 30 mins ~ 4 hrs

Daily treatment position localization/correction

5~10 mins per fraction

Weekly volume/dose evaluation

2~5 hrs per week per patient

Who should do it in long term, Physicist or RTT?

Practical Issues (workload)

Decision of Modification: Cut-off value based on

Change of patient/organ volume?

Shrinkage of the target?

Patient weight loss?

Overdosing to a critical organ?

Hot-spots on oral mucosa?

Underdosing in targets?

OR

“Expected Improvement“ of organ dose-volume obtained from the adaptive plan candidate

Practical Issues

Treatment QA

Manual target delineated on the new CT could be quite different than the auto-one. How to add dose in the target?

Missing daily CBCT image

Increased clinical QA activity & error report

Workflow management: procedure tracking & notification

Proper documentation for billing

Practical Issues

Adaptive radiotherapy of HN cancer with daily image feedback & adaptive planning modification is feasible in the routine clinic

Significant improvement in normal tissue dose could be achieved by multiple weekly replanning, or optimized by adaptive inverse planning;

Average 10% ~ 18% improvement can be achieved for most of normal organs using a single adaptive modification

Average 15% ~ 29% improvement can be achieved using the weekly adaptive modifications

The main challenge in clinical implementation is now the lack of necessary software tools, and clinical workflow support

Summary

Acknowledgement

Technology R&D:

Jian Liang, PhD

Mattias Birkner, PhD

Yuwei Chi, PhD

An Qin, PhD

David Gerstein, MS

Tie Zhang, PhD

Clinical Development:

Shanna Martin, RTT

Jenni Wloch, RTT

Qian Liu, MS

Dan Krauss, M.D.

Peter Chen, M.D.

HN ART R&D has been supported by

Elekta; Philips